Best Practices in ESG Investing
HERB BLANK, GREGG SGAMBATI, AND ZACK TRUELSON
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.
E
H ERB BLANK nvironmental, social, and corpo-
is the head of business rate governance (ESG) investing
development at S-Network
has exploded. As of January 2015,
Global Indexes, Inc., in
New York, NY. investing in ESG portfolios had
hblank@snetworkinc.com increased 76% since 2012 and accounted
for more than one-sixth of all investing
GREGG SGAMBATI (McQuillen [2015]). As the more conscien-
is the head of ESG Solu- tious approaches to investing have evolved
tions at S-Network Global
Indexes, Inc., in New
from simple avoidance of “sin stocks” (such as
York, NY. those in tobacco and alcohol) to increasingly
gsgambati@snetworkinc.com nuanced and performance-driven strategies,
demand has grown for reliable data. While
ZACK T RUELSON popular interest in such strategies has been
is a marketing associate
on the rise for decades, their focus and ter-
at S-Network Global
Indexes, Inc., in New minology have shifted away from “ethical
York, NY. investing” and toward “socially responsible
ztruelson@snetworkinc.com investing” (SRI) and similar names (Eccles
and Viviers [2011]). From the beginning, the
trend has been somewhere under the collec-
tive umbrella of ESG issues, but recently, we
have seen attention more broadly distributed
to cover the full scope of ESG. Instead of
using such a blunt instrument as negative
screening at the industry level, more investors
are paying attention to what distinguishes
companies within industries, opening an array
of new inquiries about how—rather than just
what—the companies produce.
ESG/SRI-oriented methods have been
around for several decades, dating back at
least to the launch of the Calvert Social
Investment Fund in 1982. It was not always
as data-driven as it is today. Although it is
Summer 2016
not a random process as operationalized by
Adler and Kritzman [2008]—who used
Monte Carlo simulations to argue that SRI
incurs opportunity cost—screening that
excludes and includes companies based solely
on their industries and products cannot be
expected to work as a reliably profitable
strategy. According to a thorough analysis by
Renneboog, Ter Horst, and Zhang [2008a],
as of December 2003, investors were in tran-
sition from simple negative screening to
more ESG-oriented methods. Globally, the
screen ratios were mixed, but particularly
in the United States, negative screens (97%)
dominated positive (69%) and anti-“sin”
screens (92%) outweighed environmental
(72%) and social/governance (68%), with
ethical screens comparable at 57%. (These
are the percentages of funds that employed
each type, so they sum to more than 100%.)
And not surprisingly, results were likewise
mediocre: another analysis (by the same
authors) of socially responsible (SR) funds
from 1992 to 2003 found them to under-
perform (Renneboog, Ter Horst, and Zhang
[2008b]).
More recent findings are still mixed.
Humphrey and Tan [2013] found no effect
on returns from either positive or negative
screens, while Trinks and Scholtens [2015]
found a significant opportunity cost for nega-
tive screening. Statman and Glushkov [2008]
found that negative screening of “shunned”
The Journal of Investing 103
 stocks (those involved with tobacco, gambling, alcohol,
weaponry, and nuclear power) led to lower returns while
selecting for higher ESG performance (using KLD scores)
led to higher returns. They argued that these contrary
effects largely canceled out in the world of SRI funds,
resulting in no net benefit. But their study does suggest
that a focus on ESG performance as opposed to negative
screening holds promise in terms of performance.
Much depends on how the researchers choose to
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
model the behavior of the portfolio manager. The study
by Adler and Kritzman modeled the stock selection as
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.
a randomized negative screen, essentially excluding
a given percentage of the universe of stocks without
regard to any of the stocks’ attributes. De and Clayman
[2015] also used a method of randomized portfolio cre-
ation, but did so after restricting the universe of stocks
according to Thomson Reuters/ASSET4 ESG ratings.
They found that in real-world practice, there is clearly
room for improvement. Wimmer [2013] found that
the average ESG ratings of the portfolios of SR mutual
funds persist for only two to three years—but rather
than being caused by high volatility in the ESG ratings
themselves, the short time horizon is caused by turnover
in the portfolios. In other words, it is not that highly
rated companies deteriorate and earn lower ESG ratings
after such a short time, but rather that the managers of
the SR funds are swapping out companies apparently
with little regard for what the ESG ratings are telling
them. We hope that as ESG ratings and similarly infor-
mative data gain prominence, they will become more
wisely used.
But in order to address those new inquiries on
a large scale, we need objective measures. Negative
screening of objectionable products requires very little
research; either a company produces a product that a
particular investor finds objectionable, or it does not.
Additionally, such basic product information is the
kind that companies invariably provide. But when we
start asking questions that could distinguish between
two companies making the same product, we soon run
into obscurity, opacity, and subjectivity where we need
clarity, transparency, and objectivity.
Simply put, comparable ratings enable better com-
parisons. They allow researchers to study the statistical
inf luence of ESG practices, and they allow investors
to apply the researchers’ findings in ever-more-refined
strategies. They encourage companies to be more trans-
parent in their practices, and they have the potential to
104 Best Practices in ESG Investing
encourage consumers to spend with more attention to
ethics and sustainability.
AN OBJECTIVE STANDARD
The Thomson Reuters corporate responsibility
ratings (TRCRR) are designed to serve as an objective
standard for assessing and comparing companies’ social
responsibility and are joined by a number of similarly
conceived rating systems.1 These ratings provide a quan-
titative grounding for the purposes of both research and
investing, whether aiming at the whole of ESG or at the
individual ESG components (which are referred to as
“pillars” in the TRCRR system).
As detailed by Blank [2013], the development of
the TRCRR involved hundreds of primary researchers
and a lot of careful thought, resulting in different designs
for the three ESG pillars. The goal was to take a vast
array of information and distil it into numerical rat-
ings that could be used to fairly compare companies
across and within industries and geographical regions.
Achieving that goal required numerous steps of con-
verting, weighting, and normalizing scores, as well as
qualitative decisions about how to group companies and
how to weigh the importance of various pieces of data.
In this article, we will describe the path of refine-
ment from the raw data points to the final scores, based on
the complete methodology detailed in the TRCRR Rule
Book [2014]. To make the description as clear as possible,
we will follow three examples as they evolve from qualita-
tive to quantitative, subjective to objective, raw to normal-
ized. You will gain an understanding of how the ratings
address the challenges we have discussed. Although you
will see a number of equations provided for clarity’s sake,
the procedures they represent will also be explained ver-
bally, so you will gain an intuitive grasp of the information,
even without having to parse mathematical expressions.
ASSEMBLING THE RAW DATA
The most labor-intensive part of the process is
at the beginning, when Thomson Reuters ASSET4
(henceforth ASSET4) researchers collect more than 500
data points for each of some 5,000 companies. A typical
company’s profile takes a week to assemble, drawing
from a variety of sources including stock exchange fil-
ings, annual reports, and news outlets. These data points
are sorted at two levels, first into eighteen categories,
Summer 2016
 second into three pillars—environmental, social, and
corporate governance.
We will follow one example from each of the three
pillars. The first, from the environmental pillar, asks, “Is
the company under the spotlight of the media because
of a controversy linked to biodiversity?”
We should note a couple of things about this ques-
tion and hold them in our memory for the next section,
in which we will see how the raw data are distilled into
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
KEY PERFORMANCE INDICATORS
The KPIs can be thought of as the basic units in
creating the ratings. They represent the first stage in
which the data points are roughly comparable, all con-
verted to values between zero and one. The KPIs will
be the units to which we apply various weights and
assign relative levels of importance (RLIs), and which
we group together by industry, region, both, or neither.

more than 250 key performance indicators (KPIs). First, To offer an analogy from linguistics, a KPI is to a mor-
note that this question requires a “yes” or “no” answer, pheme (the smallest meaningful unit of speech) as an
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.

which makes it a Boolean rather than a metric question individual data point is to a phoneme (a mere sound).
(which might be answered either on a 1–5 scale or with Boolean KPIs are created using one of the fol-
an absolute number). Second, note that answering “yes” lowing conversion tables (Exhibit 1 being for positive
is not a good thing in this case. That means that this ques- polarity, and Exhibit 2 being for negative polarity):
tion is one of negative polarity (if a “yes” answer were a good Answering “yes” to a one-question KPI is scored
thing, then the question would have positive polarity). the same as answering “yes” to both questions in a
The second example is from the social pillar: two-question KPI. If the company answers “yes” to one
“Does the company report or show to use human and “no” to the other, the score is 0.5. Note how “not rel-
rights criteria in the selection or monitoring process evant” (NR) and “not available” (NA) are treated differ-
of its suppliers or sourcing partners? And does the com- ently. “Not relevant” means that the question is irrelevant
pany report or show to be ready to end a partnership
with a sourcing partner if human rights criteria are EXHIBIT 1
not met?” Positive Polarity KPI
Note that this example actually consists of two
distinct questions, but they will be lumped together into
a single KPI because the second question is dependent on
the first; a company that does not pay attention to human
rights when selecting its suppliers is not likely to drop
those suppliers for human rights-related reasons. But we
would not want to discard the second question, because
follow-through is an important factor to measure. (In the
next section, we will see how partial credit is given for
“yes/no” and “no/yes” answers to these dual-question
KPIs.) Also note that both of the questions are of positive
polarity, unlike that in the first example.
The third example is from the corporate governance EXHIBIT 2
pillar, and it requires simply the “percentage of shares held Negative Polarity KPI
by all insiders and 5% owners.” Hopefully, you recognize
this as a metric question, as opposed to a Boolean one.
This metric question in particular is asking for an absolute
number rather than a number on the relative 1–5 scale,
and converting so many disparate quantities into com-
parable ones will bring us to the first of many equations
in this article (but accompanied by accessible wording,
never fear!). And thankfully, due to the way the conver-
sion works, we do not need to be concerned whether this
question is of positive or negative polarity.

Summer 2016 The Journal of Investing 105

 to the company’s region, industry, or both. For example,
the question “Is the company developing hybrid vehicles”
is relevant to the automobile industry, but decidedly not
to the telecommunications industry. A car company that
invests nothing in developing hybrid vehicles may be
fairly docked in its ESG rating, but it would be unfair
to similarly dock a telecom company, because a typical
company in that sector could not be expected to have
activity in vehicle production of any kind.
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
“Not available” is treated more stringently, as this
means that the company is simply not providing the
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.
information. We score this lack of an answer the same as
the lowest-scoring answer to that question given by any
comparable company (that is, one in the same region,
industry, or both). The logic behind this method is as
follows: If we scored NA as 0.5 like we do with NR,
then companies would have an easy way to game the
ratings. For any question to which their answer would be
negative, the company would be incentivized to refuse
to give any answer, thereby getting a 0.5 instead of a 0.
By scoring NA as the worst comparable score, we incen-
tivize companies to provide the information, but we also
do not punish them unduly for lacking it, as they can
never be the worst-scoring company as a result.
We can easily see where our environmental and
social pillar examples would fall in the previous exhibits.
Although that is quite straightforward, the metric KPIs
require some calculation to make the quantities compa-
rable. Our corporate governance example will undergo
the following procedure:
1. Subtract the lowest reported number (among
regional peers) from the company’s reported
number.
2. Subtract the lowest reported number from the
highest reported number (again, that’s the highest
reported from among companies within the same
geographical region).
3. Divide the result of (1) by the result of (2).
4. Subtract the result of (3) from one.
As you might expect, the above steps will vary
depending on the polarity of the question—if our metric
example had been of positive polarity, then we would
have skipped (4). But there is also a twist for metric KPIs
in the environmental pillar, which undergo two addi-
tional steps: the number—that is the quotient from (3),
or if applicable, the difference from (4)—is multiplied
106 Best Practices in ESG Investing
by 0.4, and then added to 0.6. This means that an envi-
ronmental KPI score cannot be lower than 0.6, which
is higher than the score given for NR (0.5). The score
given to an environmental KPI for NA is 0.4.
Alternatively, we can express the preceding in an
equation (this one being for a negative polarity environ-
mental metric KPI):
⎛ ⎛ xi min i x ⎞⎞
Score = 0.6 0.4 ∗ ⎜ 1 − ⎜ x I
⎟⎟ (1)
⎜⎝ ⎜⎝ max x min i x ⎟⎠ ⎟⎠
x I x I
where x ∈ I tells us that when we are finding the
maximum reported value x, we are considering only
the values reported by companies within the same
industry I. The symbol ∈ means “is an element of,”
so the expression as a whole is telling us that we are
taking the maximum value of x given that x is reported
by a company in the same industry as the company that
reported xi, the individual value we are converting into
a KPI score.
Why do we use the extra steps for environmental
metric KPIs? Many KPIs in that pillar have low rates
of reporting, which would result in numerous scores of
zero. To adjust for that and preserve comparability across
pillars, we set the NA score for metric KPIs at 0.4 instead
of 0 and use the additional steps for reported metrics, so
that reporting anything guarantees a company a score of
higher than NA or NR, starting at 0.6 and maxing out
at 1. The NA scores being set at 0.4 gives the distribu-
tion a more normal shape (recall that the NR score is
0.5, halfway between the NA score and the minimum
score for a reported metric).
For the positive polarity version, we would simply
remove the subtraction from one, and for an analogous
KPI in the corporate governance or social pillars, we
would furthermore remove the multiplication by 0.4 and
the addition to 0.6. To prevent any confusion, we have
provided a full list of the equations in the Appendix.
RELATIVE LEVELS OF IMPORTANCE
The second step is to assign each KPI an RLI.
RLIs range from 0 to 5 and are assigned based on sev-
eral factors:
1. How relevant that KPI is to the peer group (that
is, companies within the same industry in the
Summer 2016
 environmental pillar; the same region in the cor-
porate governance pillar; and in the social pillar,
the same industry for some KPIs, the same region
for others, and the entire universe for a few).
2. The percentage of companies in the peer group
that provide answers for the KPI.
3. The range, skewness, and standard deviation of
answers to the KPI.
4. The KPI’s independent information content.
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
5. How objectively measurable the KPI is.
6. How robustly the statistical results for the KPI can
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.
be validated using the academic literature.
As you can see, assigning the RLIs involves some
degree of subjective judgment and a great deal of simpli-
fication, as all these considerations condense into a single
value. That said, there are no subjective judgments being
made on a company-by-company basis; all of the above
considerations translate into quantitative indicators (if
they do not begin as such) and are applied consistently,
so we need not think of the RLI assignment as any
more hazardous than, for example, the two additional
mathematical steps used to convert metric KPIs in the
environmental pillar. Any arbitrariness is spread evenly
across companies, so none is especially advantaged or
disadvantaged.
CALCULATING THE WEIGHTS
Once assigned, each RLI is modified by a multi-
plier that is based on the percentage of companies in the
peer group that gave an answer for the KPI in question.
The multiplier can take one of three values: 0, 0.5, or
1. The percentage thresholds that determine the multi-
plier vary by pillar; the environmental pillar has lower
reporting percentages in general compared with the
other two, so its thresholds are adjusted downward.
• Weights in the Environmental Pillar: 0 (for when
less than 10% of the companies in the peer group
gave answers to the KPI), 0.5 (for when between
10% and 30% gave answers), or 1 (for when more
than 30% gave answers).
• Weights in the Corporate Governance and Social
Pillars: 0 (for when less than 0.5% of the companies
in the peer group gave answers to the KPI), 0.5
(for when between 0.5% and 15% gave answers),
or 1 (for when more than 15% gave answers).
Summer 2016
Next, KPI weights are determined. Uniquely in the
environmental pillar, they are split by ten factors. Because
the environmental pillar in particular has low reporting
rates, it is especially sensitive to the comparability within
the peer group (in this case the company’s industry). It
turned out that reporting percentages could be grouped
more cleanly by the factor of the KPI than by the industry
of the company. Since the divisor weight is redistributed
within the industry, the divisor needs to differ by factor
to adjust for the differences in reporting across industries
and ensure proper peer-to-peer comparisons.
Most of the factors map onto one within-pillar
category each, except for two of the factors (7 and 10)
that are represented in all three categories of the envi-
ronmental pillar (resource reduction, emission reduc-
tion, and product innovation). Exhibit 3 shows all the
factors.
To calculate an environmental KPI’s weight, we
add together all the modified RLIs of the KPIs within
a factor to get a divisor for that factor, divide the KPI’s
RLI by the factor’s divisor, and finally multiply by the
factor’s weight, as follows:
Environmental KPI
P Weiight
⎛ RLI
L ⎞
=⎜ ⎟ ∗ Factor Weight (2)
⎝ ∑ ( RLI
L Multiiplier ) ⎠
We then dynamically scale the scores on the KPI
for the peer group, such that the best score is equal to 1
and the worst score is equal to 0.
For the social and governance pillars, we likewise
add the RLIs for the KPIs in each peer group to create
divisors for the respective peer groups and divide the
EXHIBIT 3
Ten Factors of the Environmental Pillar
The Journal of Investing 107
 RLI by the peer group’s divisor, but then go straight
to the scaling step (there is no need to multiply by the
factor’s weight, since the factor divisions exist only in
the environmental pillar).2
After replicating the above process for every KPI,
we can calculate the raw score for each pillar simply by
adding together all the products of the KPI values and
their weights. In this way, all the environmental KPIs
combine into an environmental raw score, and likewise
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
with the social and corporate governance KPIs, thereby
creating three raw pillar scores that we will convert into
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.
finalized ratings in the next section.
Every KPI is benchmarked according to industry,
region, or the entire universe of companies. In other
words, a company’s score on a given KPI may alterna-
tively be compared with the scores of other companies
in the same industry or region, or simply all the com-
panies in the universe. Corporate governance KPIs are
all compared within regions; environmental KPIs are all
compared within industries; and social KPIs are variously
compared within regions, industries, and the universe.
FROM RAW SCORES TO RATINGS
This final section of methods contains the most
steps and is perhaps the most abstract. It will take us
from the raw pillar scores to the final, publishable pillar
ratings (which are combined to form the composite ESG
scores).
First, we calculate the Z-score for each raw score.
The Z-score is a way of expressing how much the raw
score deviates from the mean compared with the other
raw scores. It is simply the number of standard deviations
away from the mean that the raw score falls, with nega-
tive values meaning so far below the mean and positive
values meaning so far above the mean, as follows:
Z -score = ( R
Raw S
Score – Raw Scores
e )
/ σ ( Raw Scores
e)
(3)
Second, we adjust the distribution for outliers. If
the Z-score is less than –3, then the adjusted Z-score is
the lowest Z-score that is greater than or equal to –3.
This is mirrored on the positive side: if the Z-score is
greater than 3, then the adjusted Z-score is the highest
Z-score that is less than or equal to 3. In other words, the
Z-scores with absolute values greater than 3 are changed
to equal those that have the greatest absolute values equal
108 Best Practices in ESG Investing
to or less than 3. This means that all the outliers are
pulled inward, just to the ends of the distribution.
Third, we calculate the skewness of the distribution
of adjusted Z-scores by subtracting the mean of the adjusted
Z-scores from the given adjusted Z-score, then dividing
the difference by the standard deviation, then cubing the
quotient, then adding together all such products (one for
each adjusted Z-score), and finally dividing the sum by
the number of Z-scores. In other words, we’re using the
average cube of the Z-score of the adjusted Z-score. This
is a lot to unpack, but it essentially means that we are
trying to tell whether (and if so, to what degree) the dis-
tribution is distorted away from a perfect bell curve in
either the positive or negative direction. We need to cube
the Z-scores in order to preserve the distinction between
positive and negative skews, because merely multiplying a
negative value by itself would result in a positive value and
thereby blind us to the direction (see Equation 4):
Skew = ⎛
1 ⎞
⎝ n ⎠
n
(4)
∗ ∑ ( adjusted Z - (adjusted Z - )) / σ )
3
i =1
Fourth, we calculate the scaling divisor (for calcu-
lating the interim ratings). The purpose of the scaling
divisor is to provide a measure of the spread of the distri-
bution. In the next step, dividing by this spread controls
for the differences in the spread between pillars.
We calculate the scaling divisor by multiplying 2
by either the absolute value of the lowest- or highest-
adjusted Z-score, whichever value is greater. We then
round the product up to the next largest integer, as
follows:
Scaling (
n divisor = Ceilinng 2 max { min(
i (adjusted Z - ),
max ( adjusted Z -score )} )
(5)
where the function ceiling consists of rounding the value
up to the next largest integer.
Fifth, we calculate the interim rating for each raw
score. This step corrects for the median drift by sub-
tracting out the skewness and median Z-score from each
individual Z-score and adding 0.5 as the new median.
As mentioned previously, we also divide the isolated
Summer 2016
 difference of the individual Z-score by the scaling
divisor, which makes these interim ratings more com-
parable across pillars, despite the different degrees of
spread.
We calculate the interim rating by subtracting the
skewness from the adjusted Z-score, then subtracting
the median adjusted Z-score from the difference, then
dividing the resulting difference by the scaling divisor,
and finally adding 0.5 to the quotient, as follows:
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
r ng = 0.5
Interim ratin
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.
Adjusted Z -score − Skew − Median adjusted
d Z -score
+
Scaling
n divisor
The sixth and final step of within-pillar calcula-
tion is designed to map the ratings in each pillar onto
a normal distribution, thus reducing the clustering or
crowding of the data points. This is done by dividing
the distribution of interim ratings into four quartiles and
translating each data point into 1 in a normal distribu-
tion according to its placement relative to the limits
of its quartile. Intuitively, this process is similar to a
painter transferring an image between differently sized
and shaped surfaces with the help of a grid.
First, we find the median of the distribution and
translate that to 0.5 in the normal distribution. Then,
we find the midpoints between the median and bounds
(upper and lower) of the original distribution and trans-
late those to 0.75 and 0.25, respectively, in the normal
distribution. For each data point, we calculate the
ratio of the difference between the normal distribution
quartile bounds to the difference between the bounds
of the original quartile and multiply that ratio by the
sum of the lower bound of the normal quartile and the
difference between the interim rating and its original
quartile’s lower bound, as follows:
⎛ ⎛ U
Rating = 100 ∗ ⎜ L1 ( )∗⎜ 1
0
⎝ ⎝ U0
where L1 and U1 are the lower and upper bounds of the
normal-distribution quartile, respectively, and L 0 and
U0 are the lower and upper bounds of original quartile,
respectively. Because the second quartile’s upper bound
is set at 0.5 and the transfer formula is multiplied by 100,
the resulting normal distribution is centered at 50 (with
a range of 0–100).
Summer 2016
After this sixth step, we are finally ready to publish
the three pillar ratings. Those pillar ratings can, in turn,
be used to calculate the composite ESG rating, by mul-
tiplying each pillar rating by one-third and then adding
the three ratings together, as follows:
1 1
Composite ESG Rating = ( Environmental
E i l ) + (Social )
3 3
1
+ (Corporate Governance )
C
3
(6) (8)
APPLICATIONS OF ESG RATINGS
Now that we have seen in detail how the TRCRR
are constructed, we will brief ly explore some ways in
which they have been (and may come to be) applied.
We have used the TRCRR as the basis for a family
of indices, in order to continuously track how ESG rat-
ings correspond to the movements of domestic and
global markets. Since launching in 2013, the Thomson
Reuters Corporate Responsibility Indices (TRCRI,
developed jointly by Thomson Reuters and S-Network
Global Indexes) have served as benchmarks for the per-
formance of the top-ESG-rated companies. As such,
they strongly correlate with the ups and downs of the
broader market, but often with slight outperformance,
perhaps due to an effect of rating-based portfolio restric-
tion on risk-adjusted return, which was found by De
and Clayman [2015]. In any case, regardless of whether
the indices outperform or underperform the market at
a given time, they serve an informative role by tracking
the universe of companies from which ESG-conscious
investors will likely pick.
The Thomson Reuters Corporate Responsibility
Index family comprises 12 region- and/or pillar-spe-
L1 ⎞ ⎞ cific indices intended as benchmarks for the perfor-
L 0 ⎟⎠ ⎟⎠ mance of high-ESG-rated stocks. Exhibit 4 shows the
breakdown.
(7) As we can see, the indices are split into three
regions of markets: the United States, developed markets
excluding the United States, and Europe. They are also
split by pillar: In addition to an index of the top stocks
by overall ESG rating, each region has indices of the top
stocks by each pillar rating.
Each of the indices is created by taking the top half of
stocks (by the relevant rating) from an underlying bench-
The Journal of Investing 109
 EXHIBIT 4 correlations for the developed markets and European ver-
Thomson Reuters Corporate Responsibility Index sions were even higher.
Family Breakdown We also have recent evidence that the TRCRR can
serve to guard portfolios against risk from those compa-
nies with the worst practices. De and Clayman [2015]
used the TRCRR to study the effects of excluding the
lowest-rated companies from a portfolio. They removed
the bottom 10% of companies by ESG rating to create a
restricted universe. Then, in order to model the approach
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.

of an active manager, they assembled 40-stock portfolios

using random selection (without replacement) from the
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.

restricted and unrestricted universes (100 portfolios from

each) and compared the performance of these simulated
portfolios. Their results were striking: The restricted
portfolios outperformed the return of the unrestricted
controls in 71% of the cases simulated.
mark index published by S-Network Global Indexes: The De and Clayman produced a number of other
U.S. indices are drawn from the S-Network U.S. Large interesting findings, such as that high ESG ratings
Cap 500 Index; the Europe indices, from the S-Network strongly correlated with low volatility (a relationship
Europe 500 Index; and the Developed Markets indices, that strengthened as overall market volatility increased);
from the S-Network Developed Markets (ex-U.S.) 500 that there was a positive correlation between ESG ratings
Index (SNDMI), which is composed of the largest half and stock performance, particularly following the 2008
(by market capitalization) of the S-Network Developed financial crisis (but none significant at other times); and
International 1000 Index (SND1000). that the companies with the highest risk-adjusted returns
The TRCRI are designed to neutralize style almost always had high ESG ratings (in a methodological
exposures and portfolio growth biases. Because the reversal of the first analysis we described). There are also
screens for the indices are based on the
TRCRR, they have built-in fair com-
parison across industries and regions, E X H I B I T 5
using a best-of-class approach that avoids Performance Comparison
over- or underweighting by sector,
country, or stock type (e.g., growth vs.
value). Sector weights for each TRCRI
are neutral with respect to the relevant
benchmark index from which the stocks
are drawn.
The companies are selected through
a combination of variously weighted
market capitalizations and ESG ratings,
thus striking a balance between high
ESG performance and prominence in the
market. The results speak for themselves;
in Exhibit 5, we show an example of the
performance comparison (in this case, the
U.S.-based Large Cap Total Return Index
compared with the S&P 500 Total Return
Index). As of September 30, 2015, the cor-
relation was 0.9897, and the benchmark Sources: Bloomberg and S-Network Global Indexes, Inc.

110 Best Practices in ESG Investing Summer 2016

 questions to explore within the framework of ESG rat- Weight Calculations
ings; in an earlier article, for instance, the same authors
found that corporate governance measures were more For environmental:
strongly related to stock returns than are the E or S com- report
e ing n < 10%
% → multipl
i ier = 0
ponents of ESG scores (De and Clayman [2010]).3 This is
the sort of richly detailed research that comparable ESG 10% ≤ report
e ing n ≤ 30% → multipl
m ltipli = 0.5
i ier
ratings enable, and we hope it continues to f lourish. report
e ing n > 30%
% → multipl
i ier = 1
Environmental K
KPI Weight
CONCLUSION
⎛ ⎞
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.

RLI
L
=⎜ ⎟ ∗ Factor Weight
We have seen how the TRCRR are constructed, ⎝ ∑ ( RLI
L Multiiplier ) ⎠
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.

a few ways they have been used, and some evidence for
the effectiveness of these and similar data-driven strate- For social or corporate governance:
gies. Hopefully, this article leaves you with an intuitive report
e ing n < 0.5%
% → multipl
0 5% i ier = 0
grasp of the logic behind the process. We also hope
that you now have enhanced confidence in the quality 0.5% ≤ report
e ing n ≤ 15% → m
multipl
ltipli = 0.5
i ier
of these ratings, such that you feel confident in using report
e ing n > 15%
% → multipl
i ier = 1
them to study how ESG practices inf luence stock per- Social or Corporate Governance KPI
P Weiight
formance, stability, popular perceptions, and countless
other variables. ⎛ RLI
L ⎞
=⎜ ⎟
⎝ ∑ ( RLI
L Multiiplier ) ⎠
APPENDIX
Converting Raw Scores to Ratings
EQUATIONS
Step 1
Metric Score Conversion
Z -score = ( raw score – e ) / σ ( raw
raw scores r scores )
For environmental, positive polarity:
Step 2
⎛ xi min i x ⎞
Score = 0.6 0.4 ∗ ⎜ x I
⎟
⎜⎝ max x min i x ⎟⎠ Right tail:
x I x I
If Z-score > 3, adjusted Z-score = max(Z-score ≤ 3)
For environmental, negative polarity:
Left tail:
⎛ ⎛ xi min i x ⎞⎞ If Z-score <−3, adjusted Z-score = min(Z-score ≥−3)
Score = 0.6 0.4 ∗ ⎜ 1 − ⎜ x I
⎟⎟
⎜⎝ ⎜⎝ max x min i x ⎟⎠ ⎟⎠
x I x I Step 3

Skew = ⎛ ⎞
For social or corporate governance, positive polarity: 1
⎝ n ⎠
xi min
i x n
Score = x R
∗ ∑ ( adjusted
d Z-- (adjusted
ddjusted Z-- )) /σ )3
max x min
i x
x R x R i =1

For social or corporate governance, negative polarity: Step 4

⎛ xi min i x ⎞ n divisor = Ceiling

Scaling (
n 2 max { min(
i (adjusted Z - ),
Score = 1 − ⎜ x R
⎟
⎜⎝ max x min
x R
i x ⎟⎠
x R
max ( adjusted
d r )}
Z -score )

Summer 2016 The Journal of Investing 111

 where ceiling rounds the value up to the next largest integer. De, I., and M. Clayman “Are All Components of ESG Score
Equally Important?” NYSSA Finance Professionals’ Post ( July
2010).
Step 5
Interim rratinng = 0.5 + ——. “The Benefits of Socially Responsible Investing: An
Active Manager’s Perspective.” The Journal of Investing, Vol.
Adjusted Z -score − Skew − Median Adjusted Z -score
24, No. 4 (2015), pp. 49-72.
Scaling
n divisor
Eccles, N.S., and S. Viviers. “The Origins and Meanings of
Step 6 Names Describing Investment Practices That Integrate a Con-
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.

sideration of ESG issues in the Academic Literature.” Journal

⎛ ⎛ U L1 ⎞ ⎞
The Journal of Investing 2016.25.2:103-112. Downloaded from www.iijournals.com by CORNELL UNIVERSITY on 06/09/16.

of Business Ethics, Vol. 104, No. 3 (2011), pp. 389-402.

Rating = 100 ∗ ⎜ L1 ( )∗⎜ 1
L 0 ⎟⎠ ⎟⎠
0
⎝ ⎝ U0
Humphrey, J.E., and D.T. Tan. “Does it Really Hurt to be
Responsible?” Journal of Business Ethics, Vol. 122, No. 3 ( July
where L1 and U1 are the lower and upper bounds of the
2013), pp. 375-386.
target rectangle, respectively, and L0 and U0 are the lower
and upper bounds of the interim rectangle, respectively. McQuillen, M.J. “”ESG Portfolios: Changing Beliefs, Per-
ceptions and Goals.” The Future of Investing, ClearBridge
Composite Score Calculation Investments, 2015.

1 1
Composite ESG Rating = ( Environmental
E i l ) + (Social ) Renneboog, L., J. Ter Horst, and C. Zhang. “The Price
3 3 of Ethics and Stakeholder Governance: The Performance
1 of Socially Responsible Mutual Funds.” Journal of Corporate
+ (Corporate Governancr e) Finance, Vol. 14, No. 3 ( June 2008a), pp. 302–322.
3

——. “Socially Responsible Investments: Institutional

ENDNOTES
1
We cannot say exactly how similar in methodology,
as rating providers vary in how much of the process they
detail in publicly available fact sheets or rule books (and it
obviously impractical to publish exhaustive methods, given
the qualitative nature of the basic data points). But most ESG
rating systems are broadly similar in that they incorporate
data on ESG practices.
2
A consequence of the way the KPI weights are calcu-
lated is that all the weights for the KPIs within a peer group
(be it industry, region, or universe) sum to one.
3
Because this study was conducted in 2010, the authors’
resources were more limited; they constructed their own
ratings using data from KLD STATS (since acquired by
MSCI).
REFERENCES
Aspects, Performance, and Investor Behavior.” Journal of
Banking and Finance, Vol. 32, No. 9 (2008b), pp. 1723-1742.
Statman, M., and D. Glushkov. “The Wages of Social Respon-
sibility.” 2008 Moskowitz Prize Winner; Financial Analysts
Journal, Vol. 65, No. 4 ( July/August 2009) pp. 1-14.
“TRCRR Rule Book.” Thomson Reuters, August 14,
2014.
Trinks, P.J., and B. Scholtens. “The Opportunity Cost of
Negative Screening in Socially Responsible Investing.”
Journal of Business Ethics ( May 2015), pp. 1-16.
Wimmer, M. “ESG Persistence in Socially Responsible
Mutual Funds.” Journal of Management and Sustainability, Vol.
3, No. 1 (2013), pp. 9-15.

Adler, T., and M. Kritzman. “The Cost of Socially Respon- To order reprints of this article, please contact Dewey Palmieri
sible Investing.” The Journal of Portfolio Management, Vol. 35, at dpalmieri@ iijournals.com or 212-224-3675.
No. 1 (2008), pp. 52-56.

Blank, H. “The ESG Decision Tree.” Thomson Reuters,

June 20, 2013.

112 Best Practices in ESG Investing Summer 2016